Under Development: This portfolio is currently being built. Some links may not work yet, but all content is accurate.
Hi, I'm Karthik Ravi
Architecting Intelligent Systems
Automotive AI | Agentic Workflows | Full Stack
I build systems that bridge the gap between complex data and actionable intelligence. Currently designing AI Fault Analysis for Software-Defined Vehicles at Volvo Cars and building privacy-first apps at TrustYou.
Technical Skills
A list of technologies I work with on a daily basis.
AI & Data Engineering
Backend & Cloud
MLOps & Infrastructure
Frontend & Mobile
Work Experience
My professional track record.
Master Thesis Researcher (AI & Automotive)
Volvo Cars
Gothenburg, Sweden
- Developing an AI-powered root cause analysis system for the Central Compute Unit of modern Software-Defined Vehicles (SDV), aiming to drastically reduce time-to-diagnosis.
- Engineering pipelines to correlate and analyze gigabytes of heterogeneous log data (DLT logs, network traffic, coredumps) across 200+ quarterly fault reports.
- Implementing a Hybrid AI approach combining Anomaly Detection for log filtering and LLMs for semantic root cause analysis.
- Focusing on automating the bottleneck of manual fault tracing, transforming a process that historically took hours or months into an efficient, AI-assisted workflow.
AI & Full Stack Developer Intern
TrustYou Family AB
Stockholm, Sweden
- Architected the AI-driven insight generation pipeline, transforming raw screen-time data into actionable behavioral insights for families.
- Developed full-stack features for the mobile application, ensuring seamless integration between the intelligence backend and the user interface.
- Implemented privacy-preserving data processing workflows to derive analytics without invasive monitoring, aligning with ethical AI principles.
AI Team Lead (Industry Collaboration)
Axis Communications
Linköping, Sweden
- Architected a privacy-preserving, metadata-first Agentic AI system that ensures strict GDPR compliance by processing anonymized security logs rather than raw video feeds.
- Engineered a modular, LLM orchestration service using Dockerized FastAPI. This decoupled architecture allows for seamless hot-swapping of local models (e.g., Llama 3 to Mistral) without system downtime.
- Deployed the entire stack as a fully air-gapped, on-premise solution, ensuring total data sovereignty and low-latency inference at the edge.
- Designed a RAG pipeline to retrieve system documentation, grounding agent responses in factual data to minimize hallucinations.
Programmer Analyst
Cognizant Technology Solutions
Coimbatore, India
- Worked as a Full Stack Developer focusing on the Healthcare, transportation and logistics sectors.
- Configured and maintained scalable APIs using the .NET framework, optimizing data processing for high-volume inbound/outbound operations.
- Managed backend systems to efficiently process large volumes of healthcare data, ensuring accuracy.
- Created comprehensive API documentation (Swagger) for internal teams and external partners, reducing integration time by 30%.
Full Stack Developer Intern
Cognizant Technology Solutions
Coimbatore, India
- Led a team in designing and developing a Blood Bank Management System using ASP.NET MVC.
- Collaborated on project planning, task allocation, and code reviews to ensure timely delivery.
- Gained hands-on experience with modern web development tools and MVC architecture principles.
Full Stack Developer Intern
Conevo Technologies
Remote
- Completed comprehensive training in Full Stack Development.
- Designed and developed a Tour Booking Website using PHP and MySQL as a full-semester project for my Bachelor's degree.
- Implemented secure user authentication, database integration, and responsive UI design.
Education
My academic journey and qualifications.
Specializing in Machine Learning/Artificial Intelligence and Software Development. Relevant coursework: Artificial Intelligence, Machine Learning, Advanced AI & ML, Advanced Web Programming, Distributed Systems.
Graduated with First Class. Specialized in Machine Learning/Artificial Intelligence, Software Development and Database Management.
Featured Projects
Showcasing my work in Full Stack Development and Local AI.
Situational Awareness Agent (Axis Communications)
An air-gapped, on-premise Agentic AI system designed for security metadata analysis. Built on a Dockerized, LLM-agnostic microservice architecture (FastAPI) to allow hot-swapping of local models. Features a privacy-first, metadata-only approach for strict GDPR compliance.
NBA Team Performance Predictor & Advanced Metrics Analysis
A comprehensive sports analytics framework designed to predict NBA game outcomes by integrating raw box scores with advanced sabermetric indicators (e.g., Four Factors, Pace). Currently expanding the ETL pipeline to ingest 5+ years of historical data and engineering lag-based features (rolling averages, fatigue metrics) to strictly prevent look-ahead bias. The project aims to benchmark the predictive performance of baseline Logistic Regression against gradient boosting ensembles (XGBoost) and sequential models.
OptiToken - LLM Serialization Benchmark
A comparative research framework evaluating the efficiency of Token-Oriented Object Notation (TOON) versus JSON in LLM contexts. Currently engineering a benchmarking pipeline to quantify token reduction rates (cost efficiency) and semantic comprehension trade-offs (accuracy) across local (Llama-3 via Ollama) and cloud (GPT-4o) environments.
FastTalk - Real-time Speech AI
A high-performance speech-to-speech pipeline (STT -> LLM -> TTS) built for real-time interaction. Currently leading the TTS module development, integrating and benchmarking open-source tools like Orpheus, Piper, and Kokoro to achieve sub-second latency.
InsightCV - AI Resume Reviewer
A fully local resume analysis tool. It extracts text from PDF/DOCX and uses a local LLaMA 3.2 model via Ollama to provide intelligent feedback offline, ensuring total data privacy.
Secure Local LLM API
A secure FastAPI wrapper converting local LLMs into REST endpoints with API key authentication and rate limiting. Allows interaction with powerful models without third-party dependencies.
RebootHub - IT Gear Marketplace
Full-stack web app for buying/selling second-hand IT gear. Features user authentication, real-time listings, and chat functionality.
Heart Disease Prediction
Machine Learning model trained on Kaggle datasets to predict heart disease risk. Validated using SVM, KNN, and Decision Trees with high accuracy metrics.
Blood Bank Management System
A web portal connecting donors with hospitals. Features email notifications, ID-based access, and donor registration to minimize manual data entry.
Local LLM Chatbot (Django)
A secure, fully local AI chatbot interface powered by Llama 3.2 and Django 5.x. Provides a ChatGPT-like experience running entirely on your own hardware.
Let's Connect
I am currently looking for full-time opportunities in Software Development and Machine Learning. Whether you have a question or just want to say hi, my inbox is always open!
© 2026 Karthik Ravi. All rights reserved. Built with Next.js & Tailwind.